• Like
  • Save
Remote sensing: a tool to monitor and assess desertification. Les dossiers thématiques du CSFD. Issue 2. 44 pp.
Upcoming SlideShare
Loading in...5
×
 

Remote sensing: a tool to monitor and assess desertification. Les dossiers thématiques du CSFD. Issue 2. 44 pp.

on

  • 1,576 views

Begni Gérard, Escadafal Richard, Fontannaz Delphine and Hong-Nga Nguyen Anne-Thérèse, 2005. Remote sensing: a tool to monitor and assess desertification. Les dossiers thématiques du CSFD. Issue 2. ...

Begni Gérard, Escadafal Richard, Fontannaz Delphine and Hong-Nga Nguyen Anne-Thérèse, 2005. Remote sensing: a tool to monitor and assess desertification. Les dossiers thématiques du CSFD. Issue 2. 44 pp. - A whole range of satellites and sensors allows environmental monitoring, comparisons in time and space, and then a better understanding of ecosystems and planet functioning. This document explains how to pass from satellite data to useful information for combating desertification.

Statistics

Views

Total Views
1,576
Views on SlideShare
1,576
Embed Views
0

Actions

Likes
1
Downloads
68
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Remote sensing: a tool to monitor and assess desertification. Les dossiers thématiques du CSFD. Issue 2. 44 pp. Remote sensing: a tool to monitor and assess desertification. Les dossiers thématiques du CSFD. Issue 2. 44 pp. Document Transcript

    • CSFD SF Les dossiers thématiques Issue 2R em o t e s e n sing : se si ng: a tool to monitor and assess desertification Comité Scientifique Français de la Désertification French Scientific Committee on Desertification
    • Les dossiers thématiques du CSFD Issue 2 Managing Editor Marc Bied-Charreton President of CSFD Emeritus Professor of the University of Versailles Saint-Quentin-en-Yvelines (UVSQ) Researcher at C3ED-UMR IRD/UVSQ (Centre of Economics and Ethics for Environment and Development) Authors Gérard Begni Director of Médias-France gerard.begni@medias.cnes.fr The French Scientific Committee on Desertification Richard Escadafal Researcher at CESBIO The creation in 1997 of the French Scientific Committee on (Centre for the Study of the Biosphere from Space) for IRD (Institut de recherche pour le développement) Desertification (CSFD) has met two concerns of the Ministries in richard.escadafal@cesbio.cnes.fr charge of the United Nations Convention to Combat Desertification. Delphine Fontannaz First, CSFD materialises the will to involve the French scientific Research Engineer at Médias-France community versed in desertification, land degradation, and delphine.fontannaz@medias.cnes.fr Anne-Thérèse Hong-Nga Nguyen development of arid, semi-arid and sub-humid areas, in generating Research Engineer at Médias-France knowledge as well as guiding and advising the policy makers and anne-therese.nguyen@medias.cnes.fr actors associated in this combat. Its other aim is to strengthen the position of this French community within the international context. Contributors In order to meet such expectations, CSFD is meant to be a driving Taoufiq Bennouna, Scientific and Technical force regarding analysis and assessment, prediction and monitoring, Adviser at OSS (Sahara and Sahel Observatory) information and promotion. Within French delegations, CSFD also Antoine Cornet, Research Manager at IRD takes part in the various statutory meetings of the organs of theÉric Delaitre, Researcher at the ROSELT/OSS (Long-TermEcological Monitoring Observatories Network / Sahara and United Nations Convention to Combat Desertification: Sahel Observatory) Regional Coordination Unit of IRD Conference of the Parties (CoP), Committee on Science and Frédéric Dumay, Design Engineer at the Laboratory Technology (CST), Committee for the Review of the Implementation of Zonal Geography for the Development (LGZD) of the Convention. It also participates in meetings of European and at the University of Reims Champagne-ArdenneMonique Mainguet, Member of the University Institute international scope. of France (IUF) and Director of the Laboratory of Zonal Geography for the Development (LGZD) CSFD includes a score of members and a President, who are appointedBernard Toutain, Researcher at Cirad Emvt (Department intuitu personae by the Minister for Research, and come from variousof Animal Production and Veterinary Medicine of the FrenchAgricultural Research Centre for International Development) specialities of the main relevant institutions and universities. CSFD is managed and hosted by the Agropolis Association that Editor gathers, in the French town of Montpellier and Languedoc-Roussillon region, a large scientific community specialised in agriculture, food Isabelle Amsallem (Agropolis Productions) and environment of tropical and Mediterranean countries. The Committee acts as an independent advisory organ; Photography credits it has neither decision-making powers nor legal status. Its operating budget is financed by subsidies from the French Ministries of Foreign Affairs and for Ecology and Sustainable Danièle Cavanna (INDIGO picture library of IRD), Development. CSFD members participate voluntarily to its activities, June Cools (EOWorks), Jean-Marc D’Herbès and Éric Delaitre (IRD/ROSELT/OSS Programme), as a contribution from the Ministry for ResearchFrédéric Dumay (LGZD), Nadia Imbert-Vier (European Space Agency), Josef Jansa and Klaus Scipal (Institute More about CSFD:of Photogrammetry and Remote Sensing, Vienna University of Technology), Lionel Jarlan (Météo-France), www.csf-desertification.org Sandrine Jauffret (OSS), Édouard Lefloc’h andet Christian Floret (CEFE – Centre for Evolutionary and Functional Ecology), Marc Leroy (Médias-France), Monique Mainguet (LGZD), Éric Mougin (CESBIO), James P. Verdin and James Rowland (U.S. Geological Survey), Seydou Traoré (AGRHYMET Regional Center), Ann Tubbeckx (VITO), Nancy Walker (W.H. Freemanand Company), Professor E. Zakarin (National Center for Radio Electronics and Communications), the POSTEL team (Pôle dObservation des Surfaces continentales par TELEdétection) as well as the authors Redaction, production and distribution of Les dossiers thématiques du CSFD are Redaction, production and distribution of Les dossiers thématiques du CSFD are of the pictures shown in this report. fully supported by this Committee through the backing of relevant fully supported by this Committee through the backing of relevant French Ministries. Les dossiers thématiques du CSFD may be freely downloaded French Ministries. Les dossiers thématiques du CSFD may be freely downloaded Design and production from the Committee website. from the Committee website. Olivier Piau (Agropolis Productions) With special contribution from With special contribution from the French Space Agency (Centre National d’Études Spatiales, CNES) the French Space Agency (Centre National d’Études Spatiales, CNES) Printed by Les Petites Affiches (Montpellier, France) Translated by Catherine Tiné Registration of copyright: on publication • ISSN: 1772-6964 1,500 copies • also available in French © CSFD/Agropolis, December 2005
    • Foreword M Marc Bied-Charreton ankind is facing a world-wide concern, i.e., President of CSFD desertification, which is both a natural phe- Emeritus Professor of the University of Versailles nomenon and a process induced by human Saint-Quentin-en-Yvelines (UVSQ) activities. Our planet and natural ecosys- Researcher at C3ED-UMR IRD/UVSQ tems have never been so much degraded by our presen- (Centre of Economics and Ethics for Environment and ce. Long considered as a local problem, desertification Development) now belongs to global issues that affect us all, whether a scientist, a decision-maker, a citizen from the South or from the North. Within such a context, it is urgent to mobi- lise the civil society and induce it to get involved. To start with, people must be given the elements necessary to understand better the desertification phenomenon and its stakes. Scientific knowledge must be brought within everyone’s reach, in a language understood by the great majority. Within this scope, the French Scientific Committee on Desertification has decided to launch a new series entitled "Les dossiers thématiques du CSFD", whose purpose is to provide appropriate scientific information on desertification, its implications and sta- kes. This series is intended for policy makers and their advisers, whether from the North or from the South, but also for the general public and for the scientific journa- lists involved in development and environment. It also aims at providing teachers, trainers and trainees with addi- tional information on various fields. Lastly, it endeavours to help spreading knowledge to the actors part of the com- bat against desertification, land degradation, and pover- ty, such as representatives of professional, non- governmental, and international solidarity organisations. A dozen reports are devoted to different themes such as biodiversity, climate change, pastoralism, remote sensing, etc; in order to take stock of the current kno- wledge on these various subjects. The goal is also to set out ideological and new concept debates, including controversial issues; to expound widely used methodo- logies and results derived from a number of projects; and lastly, to supply operational and intellectual references, addresses and useful websites. These reports are to be broadly circulated, especially within the countries most affected by desertification, by e-mail (upon request), through our website, and in print. Your feedback and suggestions will be much appreciated! Redaction, production and distribution of "Les dossiers thématiques du CSFD" are fully supported by this Committee thanks to the backing of relevant French Ministries. The opinions expressed in these reports are endorsed by the Committee. 1
    • Preamble S Hubert Curien ince the dawn of time, the first hunters, and later Member of the French the first shepherds and farmers, observed their Academy of Sciences environment with their eyes and brain. They thus conceived systems of interpretation that enabled them to know where to sow and plant, where to graze their animals, and where to build their villages. Then several big revolutions occurred, in particular during the 19th century: opticians invented telescopes and binoculars, Niepce and Daguerre invented photography, and the brilliant Nadar was the first ever to set up a photographic camera in the gondola of a balloon. Aerial photography was born. Initially much used during World War I to locate the ene- mys position, this technique expanded out of the milita- ry field to become the essential tool of every cartographer and town and country planner around the world. At the beginning of the 60, the first meteorological satellites appeared, which have become essential to short-term forecasts. Earth observation satellites came out in 1972 with the US Landsat series, and the generation of high resolution satellites began with the French SPOT satellite in 1986. Today, a wide range of high, medium and low resolution satellites and sensors are available to monitor our environment, to make comparisons in time and space, and to model our ecosystems and planet in order to know them better. Thanks to these means, lots of data are received daily, but they are too often the privilege of scientists and peoplein memoriam highly skilled in their processing. Their use in developing countries, and especially in arid, semi-arid and sub-humid Hubert Curien redacted the areas, began about two decades ago. The huge services preamble of this report in January that these new techniques could supply, in particular to 2005, little before departing this life assess degraded areas and try to forecast trends, were soon on February 6th. We wish to pay him realised. a special tribute, considering his Considering that such techniques should not remain the constant concern to disseminate prerogative of technicians of developed countries, many scientific and technological results cooperative actions have been implemented and are still to the greatest possible majority. ongoing within bilateral or international frameworks. To allow development stakeholders and decision-makers toWe will miss a great man who had been General use the results obtained, it is necessary to popularise suchDirector of the French National Centre for Scientific results and to provide information regarding their limitsResearch (CNRS), President of the French Space and costs.Agency (Centre National d’Études Spatiales, CNES), This is the aim of the current CSFD publication, and IMinister of Research, President of the French Academy congratulate the Committee and the authors on theirof Sciences and President of the Cirad Ethics efforts in making accessible to a wide audience the com-Committee. plex steps required to convert data recorded onboard satel- lites into useful information.2 Remote sensing, a tool to monitor and assess desertification
    • Table of Contents4 30Remote sensing, a support Remote sensing: successes,to the study and monitoring of limits, open questionsthe Earth environment and outlook12Remote sensing applied 33to desertification monitoring List of acronyms and abbreviations22A few examples of remote 34sensing used at various scales To go deeper… 3
    • Re mote sensing, a supportto the study and monitoringof the Earth environmentE arth observation technologies play a major part in the study, modelling and monitoring of envi- ronmental phenomena, at various spatial and temporal scales, and on an objective, exhausti- ve and permanent basis. These technologies therefore open the way for the implementation of early warning systems, and capacitate policy- and decision-makers to set out relevant strategies for sustainable development. Various national and international programmes for space- based Earth observation (LANDSAT, SPOT, IRS, ERS, ADEOS, RADARSAT, ENVISAT, TERRA, METEOSAT, MSG, etc.) have been implemented as soon as 1960 and are still ongoing. They evidence the degree of priority that States (among which France plays a prominent part) attach to this technology. Up to now, the progress achieved (satellite design, measurement instruments, etc.) offers increasing capabilities to study and monitor our envi- ronment as well as global change. Remote sensing: both a science and a technology Remote sensing is defined as “all the knowledge and tech- niques used to determine the physical and biological cha- racteristics of an object by measuring it without physical contact with it” (From Journal Officiel1, December 11th, 1980). Instead of this quite broad definition, remote sensing is usually understood as a tool that allows to study phenomena involving only electromagnetic waves, Global view of the Earth © NASA-MODIS mainly detected and recorded by sensors onboard Source: NASA “Visible Earth” website, planes or satellites2. Remote sensing is consequently a consulted on January 11th, 2005. http://visibleearth.nasa.gov way to define an object or group of objects on the Earth surface from its particular features: • A spectral signature, i.e. a characteristic electroma- gnetic signal or set of electromagnetic signals in specific more or less narrow wavelength(s) of the electromagnetic Remote sensing is both a technology and a science that range; enables to observe and analyse our environment and sub- • A temporal variation in this spectral signature; sequently to define, monitor and assess policies for natural • A determined spatial distribution of this object; resource management. Satellite-based remote sensing is • One or several relations of this object with the other currently one of the only tools that allow to collect objects that surround it, i.e. the so-called “neighbouring detailed information (quite) anywhere on Earth, quickly objects”. and objectively, regularly and repetitively, thus enabling to monitor environmental events (pollution, forest fires, earth- Vegetation, lands, rivers, water-covered areas, buildings, quakes, floods, desertification, etc.). It also allows to derive and generally speaking, any element located on the Earth applications in many fields such as agriculture, forestry, surface and interacting with an electromagnetic hydrology and water resources, oceanology, geology, map- radiation, are considered as objects. ping, town planning, cadastre, as well as strategic information (most of remote sensing techniques were first 1Bulletin of official announcements issued by the French Republic (translator’s note). 2 Thisuse is not universal; for instance, our Russian colleagues have a much developed for military purposes). wider definition of remote sensing. 4 Remote sensing, a tool to monitor and assess desertification
    • Aerospace remote sensing appeared in the 60’s, but was really developed at an international scale with the NASA (National Aeronautics and Space Administration) LANDSAT programme in 1972. A second key date was indisputably the launching of the SPOT satellite by France (with Swedish and Belgian contributions) in 1986. A number of programmes and satellites have followed since then, and the design of satellites together with the conception and variety of measurement instruments have been substantially improved, thus enabling to collect a great variety of highly accurate top-quality data. Space- based remote sensing already offers considerable capa- bilities, but many studies remain to be undertaken in order to further enhance its use.Brief history Remote sensing principles: the basicsRemote sensing was born with the first aerial black and Remote sensing uses the physical properties of objects,white photograph taken by Nadar from a balloon above commonly called targets, to collect information on theirParis in 1858. However, aerial photography, that allows to nature and define them. It supposes an interaction bet-obtain a global vision of our environment, was actually ween the energy that is transmitted by electromagneticdeveloped during World War I. At first limited to the visi- radiation coming from a natural (e.g. the sun) or artificialble range (wavelength [λ] between violet [0.4 µm] and (e.g. microwave emission) source, and the target. Thisred [0.8 µm]), photography was then extended to the near energy is then sensed by an observing system, the sensorinfrared radiation (λ between 0.8 µm and 1 µm). From the (embarked onboard a satellite), that records it and trans-60’s, its until then military use was broadened to civilian mits it to a receiving station, then transforming this signalapplications such as vegetation study. into a digital image. Electromagnetic radiation interactsFrom World War II, airborne remote sensing techniques with the atmosphere a first time when it passes throughwere enhanced, especially through the development of from the source to the target, and then in the oppositenew instruments such as radars (the first imaging radars direction, from the target to the sensor. These interactionswere made in England in order to improve the accuracy induce modifications in the electromagnetic signal, whichof night bombing). are used to characterise the object observed at ground.Remote sensing, a support to the study and monitoring of the Earth environment 5
    • Basic physical foundations• Electromagnetic spectrum and radiation sourcesThe electromagnetic spectrum is divided into differentranges, from short to long wavelengths. Space-basedremote sensing only uses part of the electromagnetic spec-trum, on technological grounds and also because theatmosphere is not “translucent” in all the wavelengths.These ranges are mainly the following ones: visible(λ between 0.4 µm and 0.8 µm), near infrared (λ between0.8 µm and 1.1 µm), middle infrared3, thermal infrared(λ between 10 µm and 12 µm, which is the radiation emit-ted under the form of heat by the Earth surface), andmicrowave range (radar remote sensing). There are two Artist view of an ENVISAT satellite © ESA-DENMAN Productionsmain types of remote sensing, passive remote sensingand active remote sensing (radar). Passive remotesensing resorts to passive sensors that measure thenatural radiation reflected by objects on the Earth The three types of electromagnetic radiation:surface, whereas with active remote sensing, the system reflected, emitted and backscattered radiationboth emits and receives an electromagnetic signal. • Signal reflected by objects on the Earth surfaceElectromagnetic radiation may be transmitted by When solar radiation hits the ground, it is in part reflecteddifferent sources: to the atmosphere off the Earth surface and objects at ground. Signal reflection depends on the nature and• The sun (visible, near and middle infrared ranges): properties of the surface and on its wavelength. Onsensors record the solar energy reflected by objects on the perfectly smooth surfaces, the whole solar energy isEarth surface. reflected in a single direction, whereas on rough surfaces,• The ground (thermal and microwave fields): remote it is reflected in every direction (as usually occurs). In suchsensing receivers record the energy emitted by the Earth case, the solar flux reflected mainly corresponds to thefrom its surface temperature. visible and near infrared ranges. Recordings are only pos-• A so-called artificial source, i.e. an active sensor sible during daytime and if the atmospheric transmission(e.g. lasers and microwave radars). of electromagnetic radiation is good. With active remote sensing (radar), the energy reflected in the direction of• Disturbances caused by the atmosphere on radiation the sensor is said to be backscattered.Solar radiation, emitted or backscattered by objects atground, is subjected to alterations or disturbances • Energy emitted by objects(refraction, absorption, scattering, proper emis- With passive remote sensing, sensors measure the energysion) of various kinds when its passes through the atmo- directly emitted by objects, in the thermal infrared as well assphere. Indeed, the atmosphere allows electromagnetic microwave ranges. This energy is related to the temperatureradiation to pass through in specific spectral bands only, and surface state of objects. Contrary to the former case, thethe so-called “atmospheric windows”. Atmospheric signal emitted can be measured night and day.influence must therefore be considered by modelling, inorder to compute fluxes measured by space-based • Energy backscattered by objectssensors4. This concerns active remote sensing: the observing system includes both a transmitter (artificial source) and a receiver, usually located at the same place. The electro-3 Middle infrared is usully limited to a wavelength range in which thermal magnetic radiation that is emitted in the direction of theemission is not significant (λ < 5 µm).4 A computer displays each digital value of an image as light intensity. target interacts with its surface and is scattered in every direction. Part of the energy is consequently reflected in6 Remote sensing, a tool to monitor and assess desertification
    • Radiation – atmosphere – target – interactionsthe direction of the sensor: this is the backscattered signal. • SensorsThe basic principle of a radar is transmission and recep- They are measurement instruments that allow to collecttion of pulses, which makes it sunlight-independent. and record data on objects observed on the Earth surfaceRadars consequently allow night and day recordings, and (in one or several given wavelengths) and to transmit themare particularly useful in cloudy areas (microwaves do not to a receiving system. There are passive sensors that onlydepend on meteorological conditions), where it is often record the solar radiation reflected or the own radiationdifficult to collect data in the visible or near infrared emitted by objects, and active sensors that both emit andranges. receive the energy reflected by the target. Sensors are characterised by their:Elements of remote sensing systems • Spatial resolution: It corresponds to the size of the smallest element (Pixel) detectable on the Earth surface.A remote sensing system is a whole combination that The sharpness and details that can be distinguished in aincludes a platform, one or several sensors, and various remotely sensed image are a function of spatial resolution.means of controlling the system and of processing the • Spectral resolution: It is defined as the width or wave-data collected. length range of the part of the electromagnetic spectrum the sensor can record and the number of channels the• Platforms sensor uses.They are aerial (plane or balloon) or spatial (satellite) vehi- • Ground swath: It is the surface observed at groundcles that embark tools (sensors) to measure and record (the targeted scene).data collected on objects observed at ground. A satellitemay be sun-synchronous or geostationary. Because of their high altitude (36,000 km), sensors of geostationary satellites observing large surfaces cannot supply detailed images of our planet. On the contrary, sensors onboard lower orbiting satellites (for instance, sun-synchronous satellites, from 750 km to 900 km height) provide detailed images but on smaller areas. • Control and receiving facilities A satellite remote sensing system is always associated with a mission (or programming) centre that regularly defines the tasks to be performed by the satellite, a control centre to pilot the satellite, data receiving and recording stations, one (or several) data pre-processing centre(s), and structures for data dissemination (distribution / Sun-synchronous orbiting satellite marketing). Pre-processing centres (that are often (in white: ground track combined with receiving stations) supply standard and swath) products of easier use.Remote sensing, a support to the study and monitoring of the Earth environment 7
    • Characteristics of the main current and future operational sensors and satellites – A non-exhaustive list Satellite Panchromatic Spatial Spectral Ground Derived products Multiband resolution resolution Swath (non-exhaustive list) Very high spatial resolution satellites SPOT 5 Panchromatic 2.5 m and 5 m Optical 60*60 km Maps (geological, soil, land cover, vulnerability Multiband 10 m 60*120 km maps), satellite image maps, informative plans (river systems, road and railway networks), Digital Terrain Models (DTMs) IKONOS 2 Panchromatic 1m Optical 11*11 km Maps, satellite image maps, informative plans, DTMs Multiband 4m QUICKBIRD Panchromatic 0.60 and 0,7m Optical 16.5 km Maps, satellite image maps, informative plans, DTMs Multiband 2.4 and 2.88 m ORBVIEW 3 Multiband 1 and 4 m Optical 8*8 km Maps, satellite image maps, informative plans, DTMs HELIOS 2A 30 cm Optical Confidential Defence Pléiades Panchromatic 0.7 m Optical 21 km Maps, satellite image maps, informative plans, DTMs (2008-2009) Multiband 2.8 m EROS A Panchromatic 1 - 1.8 m Optical 12.5*12.5 km Maps, satellite image maps, informative plans, DTMs ROCSAT-2 Panchromatic 2-5m Optical 24*24 km Maps, satellite image maps, informative plans, DTMs Multiband 8 - 20 m IRS-P6 Multiband 5.8 m Optical 24 to 70 km Maps, satellite image maps, informative plans 23 m 140 km 60 - 70 m 740 km RADARSAT-1 3 to 100 m Radar 20 to 500 km Informative plans, DTMs, maps (soil moisture, flooded area maps) Medium spatial resolution satellites ERS 1,2 25 m Radar 100 km Coherence products that allow to derive land cover (especially in tropical areas) and geological maps. Soil moisture, flooded area maps. DTMs SPOT 1, 2, 3 Panchromatic 10 m Optical 60*60 km Maps, satellite image maps, DTMs, informative plans and 4 Multiband 20 m 60*80 km LANDSAT 7 (ETM) Panchromatic 15 m Optical 185*170 km Maps, satellite image maps, DTMs, informative plans Multiband 30 m LANDSAT 4, 5 Multiband 30 and 80 m Optical 185 km Maps, satellite image maps, informative plans ENVISAT (ASAR) 10 to 1,000 m Radar 15*5 km to Maps (geological, topographical, soil moisture, 405*405 km flooded area, marine pollution, coastal dynamics, glaciology maps), informative plans, DTMs TERRA (ASTER) Multiband 15 to 90 m Optical 60 km Maps, satellite image maps, informative plans, DTMs Low spatial resolution satellites SPOT Multiband 1 km Optical 2*2 km Synthesis products (daily, ten-day syntheses), (VEGETATION) NDVI (Normalised Difference Vegetation Index) METEOSAT Multiband 2.25 and 4.5 Optical Hemisphere Meteorological, oceanographic km and geophysical products MSG (Meteosat Multiband Optical Hemisphere Meteorological, oceanographic Second 1 and 3 km and geophysical products Generation) ENVISAT (MERIS) Multiband Optical 1,150 km Products derived from ocean colour measurements 300 m (carbon cycle, fishing area management, coastal area management…) SMOS Radar 1,000 km Maps (soil moisture, ocean salinity maps) (Feb. 2007) 35 and 50 km PARASOL Multiband Optical 2,400 km Radiative budget maps, observation 6*7 km of clouds and aerosolsCorrespondence between satellite image resolution and map scale:1,000 m -> 1/1,500,000 • 30 m -> 1/80,000 • 20 m- > 1/50,000 • 10 m- > 1/24,000 • 5 m -> 1/12,000 • 1 m -> 1/2,0008 Remote sensing, a tool to monitor and assess desertification
    • Scope of a remote sensing systemWhat is the use of remotely sensed data? (represented by a digital number). The intensity value cor- responds to the electromagnetic signal measured by theA remote sensing system is not by itself self-sufficient to sensor, in a given wavelength range (named spectralgenerate information directly usable by end-users. It is band). Multispectral sensors record this physical meas-first and foremost a tool for data production. Such data urement in various spectral bands (or channels); theirare then analysed together with other data sources (field number and type vary according to the sensor. Therefore,data, socio-economic data, etc.) in order to derive unders- several radiometric values correspond to each image pixel;tandable useful information likely to be integrated into this is the spectral signature of an object at a given time.information and decision support systems (GeographicInformation Systems). Digital images are displayed by associating each spectral band with a primary colour (red, green, blue). CombiningA remote sensing system may be used in various contexts; the various colours that correspond to the variousit especially plays an essential part within the scope of the intensity values of a given pixel generates a combinedcombat against desertification. Indeed, it allows to colour (additive colour synthesis). This is how colour com-follow-up and monitor the environment in the long term, posites are obtained; the most widespread process isto detect risk areas, to determine desertification factors, to called “false colour” by analogy with aerial infraredsupport decision-makers in defining relevant measures photographs (both use the same combination of spectralof environmental management, and to assess their bands).impacts. Images are generally pre-processed in two ways:Various steps to get satellite images geometric corrections (to comply with a system of map projection for instance) and radiometricSensors record the spectral response of objects observed corrections (that take into account the bias induced byon the Earth surface. Such data are neither photographs the atmosphere and convert space-based measurementsnor direct images. A satellite image consists of a two- into “ground” values). Images may be further processeddimensional array of points (pixels). Each pixel has in order to improve their readability or to extract specificcoordinates (a location address) and an intensity value information related to a particular study.Remote sensing, a support to the study and monitoring of the Earth environment 9
    • Artist view of an ENVISAT satellite © ESA-DENMAN ProductionsHow much do satellite images cost? Geostationary satellites (METEOSAT, MSG, GOES) obs-Prices vary a great deal according to the category and erve with a high temporal repetitivity the same area ofcharacteristics of a given image and to the different the Earth surface.suppliers’ policies. • Synthetic Aperture Radar (SAR) images are used toTwo major categories should be distinguished: archive measure the physical and geometric characteristics ofimages and programmed images. Archive images are the objects observed (structure, water content, biomass).scenes already recorded that are kept by imagery They allow various applications: floods, forest fires, vege-suppliers. They may be consulted and ordered. Since tation growth, soil moisture, ploughing, deforestation,the launch of the first SPOT satellite in 1986, the Spot etc. Such images are derived from ERS, ENVISAT andImage company has created an archive of more than 10 RADARSAT satellites.million images that may be consulted through thecompany’s SIRIUS catalogue. Programmed images are A number of satellite images are freely displayed on spe-future images to be collected according to the needs cific websites, such as the site developed by VITO (Flemishand characteristics defined by the customer. They are Institute for Technological Research) that distributes SPOT-generally more expensive; their price also depends on VEGETATION products (http://free.vgt.vito.be). However,the features of the image ordered (image resolution, to obtain a particular satellite image, it is usually neces-type of satellite): sary to apply to image providers. Such images are often expensive (from 1,600 to 13,000 euros), but several pro-• Very high resolution images allow to detect objects of a grammes of support to the scientific community, such assize ranging from tens of centimetres to metres. They are the French ISIS5 programme (Incitation à l’utilisationwidely used in defence and town planning sectors. They are Scientifique des Images SPOT) grant special prices.supplied by military satellites (including Helios, confiden-tial data) and in the civilian field, by the Quickbird and Within the scope of the ISIS programme, for instance,Ikonos commercial satellites. the price to be paid by laboratories for archive images (i.e. already recorded by SPOT satellites) range from• Medium resolution images (of about tens of metres) 100 to 400 euros, and from 500 to 800 euros forallow to classify lands and to locate and differentiate programmed images. These rates should be put intoforest covers and agricultural lands. They are provided perspective with the cost of a whole integratedby the Landsat, Spot and ERS satellites. application, and compared with prices for methodo- logies relying on different sources. Under the current• Low resolution images (ranging from hundreds to thou- economic circumstances, a well-considered use ofsand metres) are used at the regional and global scales. remote sensing often proves to be cost-effective.When regularly repeated, they are used to monitor envi-ronmental phenomena, regarding for instance vegetationcover, coastal areas and ocean surfaces. Low resolution 5 In 1999, the French Space Agency (Centre National d’Études Spatiales, CNES) decided to extend its ISIS programme to European scientists by granting them specialimages are supplied by the SPOT-VEGETATION, prices (excluding VEGETATION products). Since this initiative, researchers haveENVISAT-MERIS, TERRA-MODIS and NOAA satellites. bought more than 2,500 SPOT images (including SPOT-5 high resolution products).10 Remote sensing, a tool to monitor and assess desertification
    • Gl ossary Radiometric corrections: They correct recorded measurementsAbsorption: It is due to the various gases and particles in relation to the specific characteristics of the viewing instrumentscomposing the atmosphere that absorb the energy emitted by and to the disturbances caused by the atmosphere in the trans-electromagnetic radiation. It varies in relation to the nature and mission of the electromagnetic signal.properties of such gases, and according to spectral wavelengths. Refraction: A geometric distortion of the trajectory of electro-Active remote sensing: A remote sensing system that both magnetic waves travelling to the Earth surface, due to variations inemits and receives electromagnetic signals (radar principle). the refractive index.Energy flux: The power emitted, transported or received in the Satellite image map: A map drawn up from satellite imagesform of electromagnetic radiation. that are combined (tessellation) and geometrically corrected in a projection system and a standardised mapping division. A map-Geometric corrections: They allow, on the one hand, to ping product generated in digital form or in print.compensate for distortions due to satellite motion and on the otherhad, to transform an image so as to represent it as a plan (map Scattering: It results from the interaction of particles and moleculesprojection). (water droplets, dust, smoke, aerosols) on the incident and reflected electromagnetic radiation. It is a function of the natureGeostationary: Geostationary satellites use to have a circular, of wavelengths, and of the turbidity and thickness of the atmo-high (about 36,000 km above the equator) orbit; their position sphere crossed by radiation..allows a non-stop monitoring of our planet because of the highrepetitivity of their image collection. Meteorological (METEOSAT) Sensor: An instrument that detects and records the energy comingor telecommunication (EUTELSAT) satellites are geostationary from the targeted scene and that indicates the correspondingsatellites. measurable electric signal.Ground swath: It is the ground surface observed. According Spatial or geometric resolution: It is the size of the pixel atto the type of sensor, its width ranges from tens to hundreds of ground. Its geometric dimension conditions the size of thekilometres. smallest element detectable on the Earth surface. It may be like- ned to the distance that must separate two objects at ground forImage: A graphic representation of electromagnetic values them to be differentiated. It is the essential parameter to recognisemeasured and recorded in digital form.. objects by their shape.Map projection: A method that consists in transforming and Spectral resolution: The sensor sensitivity to specific wave-representing on a bi-dimensional (flat) surface, points located on the lengths of the electromagnetic spectrum. It characterises thetri-dimensional spherical surface of the Earth. accuracy of the radiometric measurement.Passive remote sensing: A remote sensing system that only records Spectral signature: The spectral response of an object,the energy reflected or emitted by objects on the ground surface. i.e. the amount of light energy reflected, absorbed by an object in various wavelengths.Photograph: An image recorded on a photographic film (throughthe chemical action of light or other radiation on specially Sun-synchronous: A satellite is said to be sun-synchronoussensitised material). when its orbit plane keeps the same orientation in relation to the Earth-Sun direction, and therefore receives the same light all yearPixel: Picture element. The smallest homogeneous surface round.A sun-synchronous satellite always crosses the vertical linecomponent of a recorded image. of a same spot at the same solar time. Most Earth observing satel- lites (SPOT, LANDSAT) are sun-synchronous.Platform: An aerial or spatial vehicle carrying the mounting (alsocalled itself a platform) where sensors and payload are set up. Synthetic Aperture Radar (SAR): A coherent radar system that generates high resolution remotely sensed images.Proper emission: The atmosphere sends back part of the incoming solar radiation, thus contributing to increase the Target: The surface or object observed (a major part of theelectromagnetic radiation reflected or emitted by the Earth surface. vocabulary used in remote sensing derives from its military origin).Remote sensing, a support to the study and monitoring of the Earth environment 11
    • Remote sensingapplied to desertificationmonitoringI t is an accepted fact that land degradation concerns our whole planet. Desertification is a typical process in arid and semi-arid areas. It also occurs in humid regions*, but is much less widespreadthere. When actually serious, desertification leads to anirreversible state of land degradation within a humangeneration (25 years). It is materialised by environ-mental changes, as land surfaces are affected by modifi-cations in both vegetation covers and soils. Desertificationfactors may be natural or human-induced. Once theyunderstand the processes that characterise this pheno-menon, researchers attempt to find indicators associa-ted with such factors, in order to assess the degree or pos-sible risks of desertification in a given area. Theseindicators allow to warn and help local or nationalauthorities in undertaking relevant actions of environ-mental management. Within this scope, remote sensingenables to assess such indicators (especially physical andecological ones) through “derived variables”, thusallowing to determine desertification processes. Sudanian savannah during the cold dry season - Autumnal landscape. Mali. Vincent Robert © IRD * In some tropical areas, for instance, deforestation followed by rain leads to soil loss and exposure of infertile bedrock. Five derived variables to assess desertification Focus indicators: roughness, albedo, surface temperature, soil moisture, vegetation indexDesertification stems from an anthropogenic process andparticularly affects economic production and consumption activities.The United Nations Convention to Combat Desertification, adopted The part played by remote sensing consists in convertingin Paris in 1994 and ratified ten year later by 190 countries, is a physical measurements of surfaces into information.Convention concerning both environment and development. It Remotely sensed data must therefore be calibrated anddefines the desertification process at the local and regional scales transformed into derived variables, that are used toas “land degradation in arid, semi-arid and dry sub-humid areas estimate desertification indicators. Information thus obtai-resulting from various factors, including climatic variations and human ned includes in particular surface roughness, albedo,activities”. surface temperature, vegetation cover (coverage andDesertification consequently describes an irreversible decline or phenological state) and soil moisture.destruction affecting the biological potential of lands and their capa-city to sustain or feed the populations. This process highlights the Roughness: quantifying the unevenness of a surfaceneed to improve the standard of living of the most vulnerablesocieties by long-term supporting their activities, preserving land Roughness is a parameter that allows to quantify thefertility or finding other activities that should alleviate pressure on unevenness of a surface. The more uneven the surface,lands. Desertification is an integral part of the issue of sustainable the greater is the roughness parameter. Roughness isdevelopment in drylands. As evidenced by the Annexes to the generally measured by radar remote sensing. Radar emitsConvention, this notion applies to every continent, mainly to dry areas micro-waves and measures the power with which anwhere aridity and drought are two common climatic data. object reflects them back (backscattering). The more une- ven the surface, the stronger is the backscattering. From Requier-Desjardins and Caron, 2005. 12 Remote sensing, a tool to monitor and assess desertification
    • ExampleUse of the “roughness” parameterThe Earth seen from a radar altimeter:land topography and ocean bathymetry © ESA/ERS This map is drawn out from data derived from the ESA (European Space Agency)ERS-2 European satellite. See original picture in the colour supplement. Example Relationships between albedo and vegetation in MaliAlbedo, i.e. the fraction of solar energyreflected from the Earth back into the atmosphere The two curves below show albedo fluctuations in the Sahel (Mali). Data are derived from the SPOT-VEGETATION satellite and concernAlbedo is the ratio of the amount of light reflected by an two spots 400 kilometres away from one another. They illustrate aobject in relation to the amount of light that hits such object, bare ground and a vegetated ground. In the former case, albedoi.e. the ratio of reflected light to incident light. Albedo is decreases around July, during the rainy season. In the latter case,expressed by a number ranging from 0 (none of the light visible albedo decreases in July whereas NIR (near infrared) albedohitting the surface is reflected) to 1 (all the light hitting the remains steady. This is due to the development of vegetation duringsurface is reflected), or by a percentage. Albedo plays a part the rainy season, which absorbs visible radiation for photosynthesis.in energetic balances and in the radiative budget since itcontrols the amount of solar energy sent back into theatmosphere.For a same geographic area, albedo may vary during theyear because of physical phenomena (or biases such asclouds in low resolution pictures). Interpreting this valuewith its temporal and spatial variations and consideringthem jointly with other observable variables supplyinformation on desertification processes. Indeed, the albe-do of a bare soil decreases when its water content in-creases. The albedo of a vegetated ground depends on itsrate of vegetation coverage and its chlorophyllous activity. Albedo fluctuation in the Sahel © Médias-FranceMany works have intended to study the relationships Data derived from VEGETATION1/SPOT4between albedo and desertification (mainly the relation- and VEGETATION2/SPOT5 satellites (2002-2003).ships between albedo and fluctuations in the vegetationcover of drylands, and between albedo and climate models). Source: FP5/CYCLOPES project (INRA [French National Institute for Agricultural Research], Médias-France, CNES [Centre National d’Études Spatiales], Météo-FranceSuch fluctuations have been evidenced at the continental and Noveltis). CYCLOPES is a shared-cost project (Contract n° EVG1-CT-2002-00076)scale, but the nature of their impact on climate is still being co-funded by the Directorate-General for Research of the European Commission withindiscussed. the Research and Development activities of the Environment and Sustainable Development sub-programme (5th Framework Programme).Remote sensing applied to desertification monitoring 13
    • Surface temperature variesaccording to land nature and coverSurface temperature results from energy exchanges thatoccur above and below this surface. It is thus partly connec-ted with albedo, air temperature and the efficiency ofthermal exchanges. Surface temperature is assessed bymeasuring the emitted thermal infrared radiation (wave-length comprised between 10.5 and 12.5 µm – passiveremote sensing). Its value depends on land nature and Soil moisture, a warning parametercover. Indeed, under the same conditions of light and for desertificationclimate, sandy or rocky grounds do not have the samebalance point temperature, all other things being equal. Soil moisture, i.e. surface water content, is defined by theSame applies to a bare or vegetated ground. Water condi- amount of water contained in the top ten centimetres oftions observed near the surface studied may also modify soil. It may be estimated by radar (active remotethe surface temperature. The time of observation is sensing), and is connected with surface temperature. Soilrelevant as well. Since the sun is the main warming moisture conditions exchanges with the atmospheresource, an image collected in the morning usually through the land surface energy budget (whichindicates surface temperatures less high than an image significantly differs in a dry or wet area). It also governscollected in the afternoon. Consequently, thanks to the growth of the vegetation cover (seed germination,geostationary satellites (e.g. meteorological satellites), it is emergence, root striking, etc.). It is consequently a para-possible to monitor the evolution of surface temperature meter well worth assessing for hydrology and agronomy,and therefore to characterise the local thermal inertia. as well as a warning parameter for desertification. Examples Variation in surface temperature Use of the “soil moisture” during a day according to surface type parameter in EuropeSurface temperature varies according to land nature and cover.The time of the day is also relevant, with a peak of temperatureat 2 P.M., i.e. when the sun, the main warming source, is at itszenith. Soil moisture in Europe – ERS satellite (January 2000) This soil moisture map is derived from measurements performed by the ERS satellite (January 2000) over Europe. Data are expressed comparatively, in percentage, with 0% representing dry lands and 100% very wet lands. See original picture in the colour supplement. From: Remote Sensing: Principles and Interpretation. 2/e by F.F. Sabins. Source: kindly provided by the Institute of Photogrammetry © 1987 by W.H. Freeman and Company. Used with permission. and Remote Sensing, Vienna University of Technology. 14 Remote sensing, a tool to monitor and assess desertification
    • ExampleGlobal scale vegetation indices © CNES 2004A global synthesis of vegetation indices made by EOworks from Image distributed by VITONDVI Spot-Vegetation data (03/1999). The colour rangeexpresses increasing index values, from yellow to green. See original picture in the colour supplement.Which processes do these parametersallow to observe?Desertification results from many natural and anthropo-genic interwoven processes, whose underlying factors areoften slow. Modelling such processes at various scalesaims at early warning, designing mitigation measures andassessing their efficiency. Models rely on socio-economicand physical observations made at ground level as wellas sensed from space, through the derived variables abovedescribed. Monitoring vegetation cover through three variables: green vegetation,Two major processes are especially dealt with in this tree density and biomassreport, since several of their characteristics are easilyobserved thanks to space-based remote sensing: • Green vegetation• Vegetation cover monitoring (which is essential Decreases in vegetation cover play a significant part inregarding food security and is a potential tracer of soil desertification processes. Many works are devoted to thefertility); monitoring of green vegetation, which is easily performed• Land cover and surface composition changes. by satellite through vegetation indices. Low resolution images daily collected by satellites allow to compute meanStudying these processes from satellite images requires 10-day vegetation indices, which in turn enable to detectto take into account seasonal cycles and rain events* . risk areas and the state of vegetation resources. It is then possible, if need be, to issue a warning to try to prevent* Generally speaking, a wet surface is less reflective than the same dry surface. the further degradation of these resources.Remote sensing applied to desertification monitoring 15
    • Examples Use of vegetation indices: Vegetation indices (NDVI) for the first ten days of April 2004 are a map of African drylands derived from NOAA/AVHRR satellite data. The higher the index, the more developed is the vegetation cover. Source: Data from NASA GSFC (Goddard Space Flight Center), GIMMS (Global Inventory Modelling and Mapping Studies). Map issued by the FEWS-NET (Famine Early Warning Systems Network) project of the USGS EROS (United States Geological Survey – Earth Resources Observation System) Data Centre. A research funded by USAID (U.S. Agency for International Development). See original picture in the colour supplement. Monitoring of grass biomass in the pastoral area of Burkina Faso Annual maps of grass biomass in the pastoral area of Burkina Faso in 1999 and 2000 (note the difference between the two annual situations).In the Sahel, annual grasses that make up a large part of pastoral Similar techniques may be applied to other semi-arid pastoral areasresources are the predominant vegetation cover. The AGRHYMET rich in annual plants, such as Central Asian regions.Centre (CILSS: Permanent Inter-state Committee for Drought Control in Monitoring vegetation north of the Sahara is more difficult becausethe Sahel) created in 1974 is currently monitoring grazing lands at the steppes of small low woody bushes prevail. Moreover, during aregional scale and is circulating information to national decision- large part of the year, vegetation is only weakly green, or even notmakers. This allows to determine grazing lands at risk, and if need be, at all during very dry periods. It is then quite impossible to useto issue a warning in order to reduce grazing in relevant regions and vegetation indices to estimate variations in vegetation cover.prevent their desertification. The state of pastoral resources is asses- © AGRHYMET Regional Centresed by estimating biomass from cumulated vegetation indices. These Source: AP3A (Early Warning and Agricultural Production Forecast) Project.data regarding the Sahelian region are also listed in the monthly AGRHYMET Regional Centre, Niamey, Nigerreport of the FAO (Food and Agriculture Organization of the UnitedNations) GIEWS (Global Information and Early Warning System). See original picture in the colour supplement.16 Remote sensing, a tool to monitor and assess desertification
    • Camp of Fulani herdsmen in a reg (desert pavement) invaded by the sand of the previous erg (Inchirian period, 22000 BP). North Oursi, Burkina Faso. Jean-Claude Leprun © IRD ExampleBiomass monitoringin the West Sahelian regionEstimating above-ground plant biomass resorts to meteorological datacombined with those derived from scatterometers (throughmodelling). In the Sahelian area subject of this study, biomassdistribution is significantly different in 1994 (wet year) and 1997(dry year). This method applied to the whole Sahelian belt shown inthis figure has been submitted to ground validation in the Gourmaregion (Mali). • Tree density This is a criterion used at ground level, especially in forested savannahs. Very high spatial resolution satellites now enable to detect individual trees and to monitor the evolution of a tree stand density. For instance, in arid areas of New Mexico, comparisons between current commercial images and old declassified pictures taken by military satellites have evidenced a high increase in the number of trees over a thirty-year period. A study of the same type (but using aerial photographs) has shown that tree stands where receding southwards in the West African Sahel. Nevertheless, this method cannot be applied to extensive areas because collecting and analysing high resolution images is quite costly. • Biomass global monitoring Vegetation indices derived from satellite-based optical Above-ground vegetation production in sensor images allow to monitor the development of green the West Sahelian region (-18°E/18°E and vegetation in land surfaces. Nevertheless, other methods 13.5°N/20°N) assessed by an ERS satellite scatterometer regarding at the continental scale have been generated from micro- these two contrasting years. wave measurements. Measurements made by scattero- © CESBIO 2003 meter sensors onboard ERS satellites (micro-wave range) Colour scale ranges from 0 (off-white) to 3,000 kg (dark brown) relate to the water content of land surfaces. Its variations of dry matter per hectare. measured in the African Sahel are particularly significant From Jarlan et al., 2003. between the dry and rainy seasons. These data allow to See original picture in the colour supplement. monitor seasonal variations in surface moisture and plant biomass. Remote sensing applied to desertification monitoring 17
    • Herdsman driving his cattle to a watering place, Burkina Faso. Marc Bournof © IRDModification in sand surface composition Land salinisation is also materialised by a loss in primaryand wind transportation: two processes productivity and is therefore indirectly detected throughobservable through remote sensing the process above described. This phenomenon rarely modifies the spectral signature of a soil so much that it mayThe primary production of a given environment mainly be unequivocally identified by its spectral signature.depends on rainfall and land states. Monitoring and However, there are extreme cases in which salt shows onmodelling from remotely sensed data endeavour to the surface and may therefore be observed in high orobtain significant indicators such as ecological efficiency, medium resolution images. This occurs for instance ingrowth balance and water consumption. A loss in land Central Asia where determined areas (usually former ponds)quality may thus be indirectly detected. are now totally converted into a salt crust. Such crust is due to both the drainage of dissolved salt and the fact that waterModifications in ground surface composition may be recharge does not offset evaporation any longer, and is veryobserved in various ways, more or less easily remotely easily observed trough remote sensing (white colour, highsensed, through variables such as reflectance and albedo).secondarily roughness. Soils and rocks may be covered with materials depositedSoil depletion due to unsuitable farming methods or or in transit, often transported by wind (dust, silt, sand) orovergrazing may reveal itself by slight changes in colour water (flooded areas). These materials are eroded and mayand/or albedo that should be interpreted cautiously. also expose underlying materials of different composition.A severe soil impoverishment often induces a decrease in Wind transportation plays a significant part in the deser-grass cover (grazing land or cultivated vegetation), which tification process. In its most striking form, it may lead tocomes down to observing the process above described. form or move sand dunes, visible by their shape and spec- tral signature (SPOT, LANDSAT, IRS images). More subtly,In determined areas where overgrazing is particularly wind transportation may occasion sand deposition likelysevere (especially near watering places), massive shuf- to invade or cover fields, infrastructures and houses. Thisfling may occur. This modifies the spectral signature and mechanism may therefore involve socio-economic dama-albedo (which usually increases) of the place. ge beyond desertification itself. Wind erosion can be miti-These localised areas may be quite easily identified gated by vegetation (trees, grasses) that stabilises sand local-through remote sensing, all the more that they often ly and contributes to slow down the wind by its aerodyna-show a specific morphology (paths converging to wate- mic roughness; vice versa, overgrazing and deforestationring points). may speed up wind transportation processes.18 Remote sensing, a tool to monitor and assess desertification
    • In this context, remote sensing is used at various scales. Aglobal picture evidences the major wind currents thatconnect for instance the Sahara with the Sahel (exampleof Sahara-Sahel wind circulation). At the regional and localscales, it allows to observe sand invasion mechanisms like- may be inferred. Low resolution satellites also enable toly to cause desertification by impoverishing soil quality, as observe the transportation of materials far away from theirwell as damage affecting infrastructures and transporta- original areas. In such cases, studies of environmental modi-tion networks. Nevertheless, it is essential to analyse aerial fications mainly rely on observing textures and structures.photographs in order to complete and refine the unders- Such observations are performed visually by a researchertanding of such phenomena. trained in these techniques. The computer processing wide- ly used in remote sensing systems proves to be less relia-Wind erosion is also detected by the frequency of dust winds ble for the monitoring of environmental degradation in drythat can be observed by systems such as METEOSAT, (sandy) regions. Whether due to erosion or sand accumu-NOAA-AVHRR and VEGETATION. Since dust winds entail lation, dunes have the same spectral signatures, anda loss in fine matter, a characterisation of soil degradation therefore cannot be differentiated by digital analyses. ExampleSpace-based observation of wind processesand sand invasion in Mauritania1- Railway of the “Société Nationale Industrie Minière” Because of the continuous sand deposition on the railway, rails mining company (Nouadhibou-Zouerate) wear out quickly: when they are not replaced in time,2- Railway section threatened by sand invasion spectacular derailments occur.3- Area of wind transit by saltation; sand veils and barchan-type structures*4- Barchan dune*: its anticlockwise tip is longer Legend: due to the coastal wind current.5- Barchan dune*: its clockwise tip is longer • Left picture: View from space: meeting place between the N/S oceanic cur- rent and NE/SW harmattan, Mauritania due to the harmattan. (from SPOT-1 P 021-313 image, 1:100,000). Kindly supplied by M. Mainguet and F. Dumay (LGZD - Laboratory of Zonal Geography for the Development). • Right picture: View at ground level: Barchan-type deposition threatening*Editor’s note: A barchan dune is a free crescent-shaped the Nouadhibou-Zouerate railway, Mauritania. © Frédéric Dumaysand dune whose crests point downwind.Remote sensing applied to desertification monitoring 19
    • Examples Detection of wind erosion by observing dust and sand winds Thanks to its 1:26,000,000 scale, this infrared METEOSAT-4 image taken on January 3rd, 1992 (METEO-France CMS Lannion) allows the global scale observation of a dust and sand storm that stretches over more than 3,000 km, from the Qattara depression (27°N, Egypt) to the Gulf of Guinea (5°N) at the limit between forest and savannah. The connections that ensure continuity between the big Saharan and Sahelian ergs have been mapped from reflectance values, just like the role played by the various obstacles on the wind trajectory of sand particles. From Mainguet and Dumay, 1995. Detection of erosion and wind transportation: the Aral sea This figure illustrates the particularly serious case of the regions surrounding the Aral sea (Central Asia). Excessive water consumption for irrigation purposes and outdated technologies have led to a terrible shrinking of this lake. While evaporating, the Aral sea leaves an infertile salty deposition. The albedo of this totally desert area is very easily detectable on low, medium and high resolution images (NOAA-AVHRR or SPOT-VEGETATION satellites). This deposition is prone to erosion and wind transportation: both processes spread salt particles and make the soils affected infertile. This may be interpreted as an extreme case of local salt crusting in this region. Kindly supplied by Pr. E. Zakarin, Observations through low resolution remote National Centre for Radio-Electronics and Communication sensing of the former bed, now dried-up, of the Aral sea (Central Asia) and of wind transportation of the Republic of Kazakhstan (NCREC) episodes (19/09/1998 and 09/04/2002). within the scope of a cooperation with Médias-France.20 Remote sensing, a tool to monitor and assess desertification
    • Gl ossary Indicator: A synthetic parameter used to assess environmental changes connected with desertification processes. Indicators may be quantitative or qualitative. Normalised Difference Vegetation Index (NDVI): The ratio of: (Near Infrared - Red) / (Near Infrared + Red). Tagant plateau. This plateau populated from remote antiquity is a heterogeneous combination Reflectance: The ratio of the total amount of radiation reflected of cliffs, crests and massifs intersected by alluvial by a surface to the total amount of radiation incident on the plains and defiles where oases and palm groves are located. N’Belka, Tagant plateau, Mauritania. surface (target). Jean-Jacques Lemasson © IRDRemote sensing applied to desertification monitoring 21
    • A few examplesof remote sensingused at various scalesR emote sensing is a space-based observing tool that allows the study and monitoring of spots at various scales. The present chapter illus- trates this variety with two examples: The Menzel Habib test site in pre-Saharan Tunisia has been subjected to repeated observations since 1970, especially within the framework of the CAMELEO (Changes in Arid Mediterranean Ecosystems on the Long term and Earth Observation) project (1997-2001); The ROSELT (Long-Term Ecological Monitoring Obser-vato- ries Network) project. This interdisciplinary initiative has been led at the regional circum-Saharan scale by the Institut de recherche pour le développement (IRD) within the scope of OSS (Sahara and Sahel Observatory) Wadi El Akarit located at some thirty kilometres programmes. north of Gabes, Tunisia. Jean-Pierre Roset © IRD Monitoring of a sandy steppe area: an example in pre-Saharan Tunisia Among the places where research on desertification has Space-based monitoring methods have been tested and been undertaken in Africa, the Menzel Habib spot in developed in this region, in particular within the scope of southern Tunisia has been continuously monitored since Euro-Mediterranean research programmes such as the the 70’s, especially by the French and French-speaking CAMELEO project, which is the source of the following scientific communities and their pioneers in this field. This information. site, that belongs to the observatories of the ROSELT programme, has been the subject of space-based Immature soils covering hard sedimentary rocks are loca- monitoring experiments and of a recent study on long- ted only in the mountains that border the Menzel Habib term ecological indicators. plain. Quite all the other soils in the region have developed on wind-deposited materials, peridesert loess and fine The Menzel Habib observatory sand, enriched with soluble elements originated from underlying gypsum and salt deposits. Organic matter is The Menzel Habib area, located in southern Tunisia, is quite scarce, but a small quantity due to surface biological characterised by its highly irregular annual rainfall ran- activity gives some cohesion to the sand (algal crust for ging from 100 to 200 mm and its sandy and sandy-silty instance). soils covered with a low woody steppe, typical of the arid regions of the northern Saharan border. Regarding natural vegetation, low woody bushes (cha- maephytes) prevail; annual plants develop quickly after Due to both a drought period and the cultivation of rain events that mainly occur in winter. Cultivated vegeta- plots until then used as rangelands, this region was tion consists of annual crops (barley, hard wheat) and trees affected by particularly severe desertification pheno- of poor planting density. Consequently, vegetation is green mena during the 80’s. A programme aimed at comba- only during part of the year (generally the first months), ting land degradation and sand invasion was conse- and shows on the whole a poor coverage. At surface level, quently launched. soils prevail. 22 Remote sensing, a tool to monitor and assess desertification
    • FocusLocation and annual rainfall,Menzel Habib, Tunisia Variation in annual rainfall in Menzel Habib, Tunisia (1970-2000). Over this thirty-year period, the mean annual rainfall amounts to 180 mm. Wetter years can be observed until 1980, then dryer years follow. They are interspersed with “rainy” years with rainfall exceeding 250 mm, as for instance in 1990. From Jauffret, 2001. Location map of the Menzel Habib test area, pre-Saharan Tunisia. Map from Floret and Pontanier, 1982.Research approach regardingthe monitoring of the Menzel Habib area abundance of woody perennial plants) globally affects signal intensity. Consequently, for a given soil, a denserRemotely sensed images have been used to establish steppe appears darker on satellite images (albedo-relationships between, on the one hand, the soil and related link). characteristics of surfaces observed at ground level, andon the other hand, the spectral response of the same In addition, since the local vegetation generally providessurfaces measured by satellite-based optical sensors. At little coverage, reflectance measurements are highlyground level, the ecological description of surface states influenced by soil properties, and especially by their(soil composition and organisation, phenological natu- colour. A colour index derived from measurements madere, vegetation abundance) have been associated with in the visible spectral bands was suggested: the higher itreflectance values measured by portable devices. It was is, the more coloured are the soils (e.g. sand), while lowthus demonstrated that in this region, the vegetation index values correspond to greyish soils (e.g. gypsum).was ill correlated with the global vegetation cover. The Consequently, both albedo and colour criteria have beenrate of vegetation cover (a highly relevant indicator regar- used to monitor the evolution of surface states over time,ding desertification diagnosis, which accounts for the and to determine trends.A few examples of remote sensing used at various scales 23
    • Monitoring soil and vegetation evolutionStudies bearing on a whole region require high resolution The pictures below display results obtained on four dates,images. A Landsat image series was collected over the when differences are particularly contrasted. CombiningMenzel Habib sandy plain, on the basis of one image per older Landsat MSS images with more recent Landsat TMyear. Whenever possible, pictures were taken during the ones allows to cover a 23-year period. Images are shownsame season in order to minimise reflectance differences in standard colour compositions (“false colours”), in whichcaused by variations in sun altitude. Spring images were green vegetation is displayed in red. Two photographspreferred so as to photograph vegetation when it reaches taken at ground level feature a sandy steppe in “normal”its maximum. These images were processed with geo- and degraded states, to complete this illustration.metric correction (to make them superimposable pixel Degradation due to a decrease in vegetation cover is mate-by pixel) and radiometric correction (to convert the pixel rialised in such case by the encroachment of moving sandvalues of each image into ground reflectance), in order to (sand invasion).make them comparable.FocusDegradation due to a decrease in the vegetation coverof the Menzel Habib sandy steppe, TunisiaView at ground level R. Escadafal © IRD Normal state Degraded stateSatellite images The differences evidenced correspond to modifications in surface states at ground level. Each picture covers a 30x24 km area. R. Escadafal © IRD • Landsat MSS image dated April 1976 The Menzel Habib plain in 1976: bordered in the South-East and South- West by mountains (in grey), a sandy steppe prevails (in beige in the centre). Annual crops (in red) are located at the periphery and in depressions (bright red spots). • Landsat TM image dated April 1989 MSS Image 04/76 TM Image 04/89 In 1989, drought leads to a decrease in crops and to the striking expansion of mobile sand (light yellow): the region is “desertified”. • Landsat TM image dated March 1993 In 1993, i.e. four years later, the impacts of exclosure areas clearly show as dark trapezium-shaped spots among sandy areas, while annual crops intensify (red dotted lines). • Landsat TM image dated March 1999 In 1999, mobile sand areas prove to have completely receded and TM Image 03/93 TM Image 03/99 the situation is apparently well controlled. The landscape appears to be fully parcelled out in plots of various uses, and new exclosure areas Four georeferenced and intercalibrated Landsat images have been set up in the few remaining dune areas. of the Menzel Habib area (Tunisia) (1976, 1989, 1993, 1999) Richard Escadafal © IRD See original picture in the colour supplement.24 Remote sensing, a tool to monitor and assess desertification
    • FocusSynthesis of sand invasion evolutionin the Menzel Habib area (Tunisia)between 1989 and 1999 This figure illustrates one of the syntheses obtained by analysing trends observed in images classified over five dates (between 1989 and 1999) and showing sand-invaded surfaces (unfixed moving sand, dunes, etc.). Sand-invaded areas (mobile dunes) have decreased in favour of fixed sand areas and farming lands. Satellite images thus evidenced that the environmental state did improve between 1989 and 1999. Efforts undertaken in order to fix the moving sand that had invaded the region during the previous decade were successful. Space-based monitoring allowed to quantify these impacts over large expanses. Such good results were of course also observed locally on site. See original picture in the colour supplement. Richard Escadafal © IRDMap synthesis of environmental evolutionObserving image series makes it possible to design asequence of the various surface states. However, it is not Application of this method to other contextssufficient to define long-term trends, nor to identify envi-ronmental desertification or restoration. Analyses should This example in southern Tunisia illustrates the use ofbear on long enough series and derive trend syntheses satellite images for the monitoring of a sandy steppe area.while taking into account climate variations. Desertification also causes havoc in other environments such as southern Sahara where it materialises different-In order to define such long-term trends, many attempts ly, with an increasing scarcity of woody plants forto monitor desertification focus on biomass. Instead of instance. Satellite-based monitoring should consequentlythis method, since biomass is too poor in the region stu- take into account the ecological features of the environ-died for its variations to be easily detectable, surface sta- ments followed-up and rely on the knowledge of localtes in general have to be monitored in the long run. processes. This is the necessary condition for being able to interpret changes in surface states and to identify landSuch surfaces of sparse vegetation cover require the use status: environmental degradation, stability or restora-of brightness and colour indices, as well as vegetation tion. Thanks to the increasing number of satellites andindices that help distinguish the most active covers. sensors, denser and more diverse data are collected fromThis method allows to classify each image according to a space. The current challenge then consists in using themsimple legend based on soil type and vegetation cover at best to obtain the most accurate monitoring at thedensity. The percentage covered by each category has been lowest cost. The ultimate goal is to supply the informa-monitored over time so as to determine environmental tion necessary for early warning systems, which are theevolution (stability, degradation or improvement). actual concern of the managers of the regions affected.A few examples of remote sensing used at various scales 25
    • Using satellite imagesfor wildlife monitoringWithin the scope of the Long-Term Ecological MonitoringObservatories Network (ROSELT) of the Sahara and SahelObservatory (OSS), ASTER images (NASA) have been usedin the Oued Mird observatory (Morocco) in order toimplement a wildlife monitoring system. Such systemintegrates the concepts and methods derived fromlandscape ecology and population biology (Baudat, 2003).An integrated mapping of the natural environment has Erg Chebbi, the beginning of the Great Western Erg located south-east of Erfoud.been designed to allow to work at various scales: from An example of sand encroachment:landscape to biotope, then to habitat. It is then possible dunes invading palm groves, Morocco. Claude Dejoux © IRDto follow up landscape evolution, to sample animalpopulations as per biotope and to take into account thespecific problems connected with habitats of particularspecies.For biotope mapping, a plan has been defined and astratified sampling method selected according to theanimal population studied. Integrating such data into a This area is inhabited by about 1,500 people who live onGeographic Information System (GIS) allows to perform extensive breeding and food-producing agriculturethe necessary processing and spatial operations. (cereals and henna). Out of 152 families registered, only 17 still practice nomadic breeding in the pre-Saharan ran-The Oued Mird observatory in Morocco gelands of this area.The Oued Mird observatory extends over 550 km² in sou- Drawing out a landscape mapthern Morocco (Saharan Anti-Atlas). This site is locatedat the Saharan bioclimatic stage, with warm winters (mean The set-up of observation stations takes into account alti-winter temperatures above 7°C and annual rainfall below tude and slope, if possible along toposequences. For every100 mm, Brignon and Sauvage, 1963). Altitude ranges bet- station, a vertical section has been drawn out: each oneween 637 and 1,243 metres, from the bottom of the Oued includes a systematic stratified description of environmentalMird valley to the crests of Jebel Tadrart. Plant commu- components, through 17 variables considered as ecologi-nities consist of desert steppes and wooded steppes. cal descriptors of desert environments (soil and vegetation). Two hundred and six stations that are geo-referenced byBesides, Oued Mird exemplifies a specific desertification GPS (Global Positioning System, four weeks at ground level)problem, because there are still in this region well conser- have thus been depicted in more than 400 photographs.ved Acacia raddiana woods. Such species is a key host A typology of natural environments could then be drawnfrom an ecological point of view. Together with access to out with the help of statistical analyses (factor analysis ofwater, it represents the socio-economic mainspring of correspondences associated with a hierarchic ascendinghuman presence in this territory. Indeed, these trees sup- classification): 14 classes evidence an ecological gradient,ply firewood for heating and cooking, and are grazed by from non-vegetated environments (regs and bare grounds)cattle (camels, goats). to Acacia raddiana woods.26 Remote sensing, a tool to monitor and assess desertification
    • Three landscapes are defined on the map: • A slope landscape: A long medium slope shows a mono-Two 15-m resolution (visible and near infrared) ASTER tonous aspect due to the prevailing reg and screesimages, both taken on the same date (August 18th, 2001) (environmental classes 14 and 11). Vegetation only consistswere classified according to the so-called “maximum like- of bushes along drainage axes (13). A matrix of much inter-lihood method” by using the typology derived from connected rocky environments can be identified. It isground stations. However, only a poor overall accuracy of intersected by vegetation corridors as well asthe classification was obtained (< 40%): ground classes scree-covered slopes with rocky ledges (7 and 11). Thewere totally mixed together or not even classified. This is whole combination forms a very homogeneous mosaic.due to the fact that ecological classification (ground level)and radiometric classification (satellite image) did not • A closed depression landscape: The Tafenna depression ismatch. In fact, the ecological variables that prevail in the a large oval basin with a quite flat bottom where fine mate-definition of these classes are not detected in the 15-m rial regs prevail (9 and 12), cut across by screes on the bareresolution satellite image and/or are completely mixed ground (11). Sandy-silty thalwegs are colonised by a well-up. In addition, the vegetation cover, which often does developed bushy vegetation, interspersed with shrubsnot even reach 10% in this area, could not be sensed pro- and isolated trees (10 and 5). Sand deposits (2) likely toperly. reach a considerable size cover the sides of the basin that are characterised by steep slopes with coarse screes andEventually, images have been visually interpreted rocky ledges (11 and 7). A fine reg gentle slope matrix canaccording to ground typology, in order to correct the be spotted. It is divided by coarse screes and corridors ofresults of the maximum likelihood classification: the map sandy-silty steppe and bush vegetation. Sharp sides withof landscapes and natural environments therefore results rocky ledges and screes constitute the transition with thefrom associating computer-aided classification with slope landscape; they form a circle-shaped corridor thathuman expertise. delimits the depression (orange line).A few examples of remote sensing used at various scales 27
    • Example of the houbara bustard The houbara bustard is a quite big (about 60 cm) walking bird whose distribution area extends from Saharan arid areas to central Asia. The habitat of this species consists of xerophytic steppes, with fine reg sandy-silty ground and gentle slopes interspersed with small depressions. The diet of the houbara bustard is mainly vegetarian, including insects (Collar, 1996). In Algeria, in the same biogeographical area as the region studied, Gaucher (1991) provided several indications regarding which plant species the houbara bustard eats (such as those of the genus Farsetia in Oued Mird), and regarding its foraging area (1 km/day within a minimum 400-m radius). A map of the houbara bustard habitat in the Oued Mird observatory has been derived from these criteria and the environmental map. A superimposition onto the topographic base allows for instance to plan a Landscape map, Oued Mird observatory, Morocco. © ROSELT/OSS.From Baudat, 2003. systematic research or to establish a baseline status of Lambert conformal conic projection / South Morocco II. the habitat for monitoring purposes. Besides, the result Merchich datum / 1880 Clarke ellipsoid (IGN). Map derived from two 15-m (visible/NIR) Aster images. tallies with the observation of an individual during the Pictures taken on August 18th, 2001. campaign of June 2003. J. Baudat / Mastère SILAT (GIS applied to land use planning) IRD-ROSELT/OSS, October 2003. See original picture in the colour supplement.• An alluvial valley landscape: The Oued Mird valley hasthe most heterogeneous ecology in the observatory. Thisis due to the existence of water (outfall of the slopedrainage system, near-surface groundwater) associatedwith human presence (crops). The matrix there consists ofmedium to coarse regs of terraces and alluvial cones (4) orof the association of silty deposits with very sparsevegetation cover and sandy-silty steppes (5 and 6). Theother element characteristic of this alluvial valleylandscape is the combination of heterogeneous spots ofmixed crops and palm groves (3), sand deposits and areas(2), an Acacia raddiana wooded steppe (1) and bare groundswith sealing crust and salt efflorescence (8).A landscape map may therefore be regarded as a meetingpoint for the various nature sciences, since a landscapemay be considered as a collection of biotopes (Forman andGodron, 1986; Blondel, 1995; Burel and Baudry, 2001). Thistype of mapping product is a tool for interdisciplinary workand long-term ecological monitoring. For ROSELT, defi- Location of the houbara bustard habitatning landscape units of this kind would allow to integrate (in dark orange) in Oued Mird. Modified topographic base: 1:100,000 (DFCTT, 1968).wildlife monitoring from the study of populations sampled © ROSELT/OSS. From Baudat, 2003.in the various biotopes. See original picture in the colour supplement.28 Remote sensing, a tool to monitor and assess desertification
    • Glossary Biotope: A localised region of uniform environmental conditions which constitutes the habitat of one or several plant and animal communities. Wadi Dades valley, North of Georeferencing: The process of assigning map coordinates to Boulmane, Morocco. Georges Grandin © IRD image data in order to conform to a map projection system.A few examples of remote sensing used at various scales 29
    • Re mote sensing:successes, limits,open questionsand outlook An increasingly affordable, reliable tool for data collection Satellites have the immense advantage of covering at a reasonable cost very large expanses in a repeated, homo- geneous and systematic way, which cannot be done on site. Naturally, this notion of “reasonable cost” should be clarified. When resolution increases, the image size decreases - generally together with observation frequency - while collection and processing costs usual- ly increase as well. Consequently, the regular monitoring of large surfaces at Young Fulanis, Ferlo, Senegal. a daily or weekly scale typically resorts to low resolution B. Devaux © IRD images. The dynamics observed in the evolution of such images then needs to be interpreted at a small scale. To make this easier, a few high resolution images may be used. These pictures, that are less frequently collected, concern representative sites selected according to their Coupling remotely sensed data with in situ information specific characteristics (vulnerability for instance) or to statistical rules (stratification and sampling), or else in For a start, there is a long way between highly complex view of inputing to long-term observatories networks. desertification processes, relevant indicators that allow to monitor them, variables observable through remote sens- Efforts are in progress to federate and improve in an ing, and the efficiency of methods that enable to extract integrated way research and the knowledge obtained up these variables and to derive models. Some difficulties are to now. In particular, the European Commission has taken partly constitutional (is it possible to determine the such initiatives within the scope of the 6th research fra- irreversible aspect of a degradation from remotely sensed mework programme. When a seemingly unusual information?), others may be solved by advances in evolution appears in low resolution images (early research. Such advances often consist in associating remo- warning), what is actually happening in such places can tely sensed data with data derived from other sources in be observed more precisely by programming high models that improve the use and efficiency of these various resolution images and if need be, by undertaking on-site data. The ROSELT project exemplifies a major effort to assessment. reduce this kind of gap in the circum-Saharan region. Indeed, many ecological diagnoses are based on the dyna- A follow-up through both low and high resolution ima- mics of different species, the extinction of some of them ges also allows to monitor the efficiency of restoration and appearance of new ones, and such information can- and combatting desertification measures. not be supplied by satellite (biodiversity aspect). Only specific populations may be recognised by their How to bridge the “digital gap”? particular structure, usually thanks to the phenological calendar that enables to differentiate easily annual plants Between experts in remote sensing digital techniques and from perennial vegetation, and even more easily deciduous stakeholders of the combat against desertification, there species from evergreen ones. Progress consists in is what is popularly called a “digital gap”, that scien- obtaining the most representative models based on tists, engineers, decision-makers and their advisers are remotely sensed information and ground data sources, trying to bridge. such models being designed to be used in the long run. 30 Remote sensing, a tool to monitor and assess desertification
    • Improving international aid targeting behind in the information distribution chain. The issue of training end-users in these new space technologies andThe cost of remote sensing is reasonable in view of the the question of staff retention still remain pending.advantages it offers compared with other techniques. Without trained people, investment and operating costsInvestments, operating costs and expenses for experts’ would be spent to no avail.training must be allowed for. However, countries threate-ned by desertification are poor nations that can hardly The optimal and efficient use of remote sensing withinfinance such investments and recurring costs, let alone keep the scope of the desertification combat is both success-specialised staff. Under these conditions, it is surprising to ful and limited at the social, economic and environmen-see that international aid generally concerns intellectual or tal levels. To go ahead, it is essential to make progress inmaterial investments rather than recurrent operating costs, reducing the three types of gaps above mentioned. Ifwhich might make fruitless the investments afforded by voluntary participation lacks, new knowledge might beboth parties (brain drain, obsolescent equipment). generated without any possibility to derive practical applications; or the methods applied by relevant stake-Expanding the dissemination and appropriation holders might unacceptably stagnate whereas technolo-by end-users of remote sensing techniques gical advances would allow to forge ahead and ensure the sustainable development of the most vulnerable regionsTechnology is developing faster than its use and experts and populations of our planet.know-how. Nevertheless, comparatively simple productsthat do not match the highest standards of technologicalknowledge are available; e.g. warning bulletins are nowcommon. For a start, the utilisation of this type ofproducts should be enhanced. In particular, it is essen- Glossarytial that end-users should get them and be able to inter-pret them. Thanks to the quick development of digital Digital gap: The gap regarding the access to informationcommunication techniques, their dissemination that used technologies.to be a problem a few years ago is being considerablyimproved - even though end-users may be a long wayRemote sensing: successes, limits, open questions and outlook 31
    • Young people of the Kamadjan IRD youth club are making a flora inventory conducted by Moussa Karembe (researcher at IER). Mande region, Mali. Thérèse Touré © IRD32 Remote sensing, a tool to monitor and assess desertification
    • List of acronyms and abbreviations AGRHYMET Regional Training Centre for Agrometeorology and Operational Hydrology and their applications / Centre Régional de Formation et d’Applications en Agrométéorologie et Hydrologie Opérationnelle AP3A Early Warning and Agricultural Production Forecast Project / Projet Alerte Précoce et Prévision des Productions Agricoles AVHRR Advanced Very High Resolution Radiometer C3ED Centre of Economics and Ethics for Environment and Development / Centre d’économie et d’éthique pour l’environnement et le développement CAMELEO Changes in Arid Mediterranean Ecosystems on the Long term and Earth Observation CESBIO Centre for the Study of the Biosphere from Space / Centre dÉtudes Spatiales de la BIOsphère CILSS Permanent Inter-State Committee for Drought Control in the Sahel / Comité permanent Inter États de Lutte contre la Sécheresse au Sahel Department of Animal Production and Veterinary Medicine of the French Agricultural Research Cirad Emvt Centre for International Development / Département Élevage et Médecine Vétérinaire du Centre de Coopération Internationale en Recherche Agronomique pour le Développement CNES French Space Agency / Centre National dÉtudes Spatiales CSFD French Scientific Committee on Desertification / Comité Scientifique Français de la Désertification DTM Digital Terrain Model EROS Earth Resources Observation System ESA European Space Agency ETM Enhanced Thematic Mapper FAO Food and Agriculture Organization of the United Nations FEWS-NET Famine Early Warning Systems Network GIEWS Global Information and Early Warning System GIMMS Global Inventory Modelling and Mapping Studies GIS Geographic Information System GPS Global Positioning System GSFC Goddard Space Flight Center INRA French National Institute for Agricultural Research / Institut National de la Recherche Agronomique IRD Institut de recherche pour le développement ISIS Incentive for the Scientific use of Spot Images / Incitation à lutilisation Scientifique des Images SPOT LGZD Laboratory of Zonal Geography for the Development / Laboratoire de Géographie Zonale pour le Développement Médias Regional Research Network on Global Change in the Mediterranean basin and subtropical Africa / Réseau Régional de Recherche sur le Changement Global dans le bassin MEDIterranéen et l’Afrique Sub-tropicale MERIS Medium Resolution Imaging Spectrometer Instrument MODIS Moderate Resolution Imaging Spectroradiometer MS Multispectral MSG Meteosat Second Generation MSS Multi Spectral Scanner NASA National Aeronautics and Space Administration NCREC National Centre for Radio-Electronics and Communications NDVI Normalised Difference Vegetation Index NIR Near InfraRed OSS Sahara and Sahel Observatory / Observatoire du Sahara et du Sahel POSTEL Pôle dObservation des Surfaces continentales par TÉLÉdétection ROSELT Long-Term Ecological Monitoring Observatories Network / Réseau d’Observatoires de Surveillance Écologique à Long Terme SAR Synthetic Aperture Radar SMOS Soil Moisture and Ocean Salinity UMR Joint Research Unit / Unité Mixte de Recherche USAID U.S. Agency for International Development USGS United States Geological Survey UVSQ University of Versailles Saint-Quentin-en-Yvelines / Université de Versailles Saint-Quentin-en-Yvelines VITO Flemish Institute for Technological Research 33
    • To go deeper...BibliographyRemote sensing Mainguet M., Dumay F Mahfoud A. and Hacen L., 1999. Baseline ., and Growth Indicators for Desertification in the Sahara - SahelianBonn F. and Rochon G., 1992. Précis de télédétection - Volume 1 Area of Mauritania and their Monitoring from 1953 to 1998.- Principes et Méthodes. Universités Francophones, Presses de Desertification Control Bulletin. 34: 21-30.lUniversité du Québec/AUPELF. 485 p. Requier-Desjardins M., 2002. Désertification et environnementGDTA, nd. Visualisation dimages numériques,Visualisation des mondial (biodiversité et changement climatique) : propositioncouleurs, Bases physiques de la télédétection. Cahiers dindicateurs pour un projet de développement, l’IREMLCD.pédagogiques du GDTA (Groupement pour le Développement FFEM/CFSD report, France. 60 p.de la Télédétection Aérospatiale), France. Requier-Desjardins M. and Caron P., 2005. La lutte contre laGirard M.C., 1989. Télédétection appliquée - zones tempérées et désertification : un bien public mondial environnemental ?intertropicales. Masson, Collection sciences agronomiques, Des éléments de réponse… Les dossiers thématiques duParis. 260 p. CSFD n°1. CSFD/Agropolis, Montpellier, France. 28 p.Girard M.C. and Girard C.M., 1999. Traitement des données de Tunisian exampletélédétection. Dunod Publishers, Paris. 527 p. Escadafal R., 1989. Caractérisation de la surface des sols aridesDesertification par observations de terrain et par télédétection. Applications : exemple de la région de Tataouine (Tunisie). Études et Thèses.Jarlan L., Mazzega P Mougin E., Schoenauer M., Lavenu F Marty ., ., Édition ORSTOM Publishers, Paris.G., Frison P.L. and Hiernaux P., 2003. Mapping of SahelianVegetation parameters from ERS Scatterometer data with an Escadafal R., 1994. Soil spectral properties and their relationevolution strategie algorithm. Remote Sensing of Environment. ships with environmental parameters – Examples from arid87: 72-84. regions. In: Hill J. & Mégier J. (Eds). Imaging spectrometry – a tool for environmental observations. Kluwer AcademicMainguet M. and Dumay F., 1995. Trans-Saharan Wind Flows Publishers, Dordrecht: 71-87.Observed on Meteosat 4 Satellite Image. Resources, urban, sand& wind, Desert Technology III, Oct. 15-20 1995, Lake Motosu, Escadafal R. and Megier J., 1998. CAMELEO: a concerted researchJapan. Journal of Arid Land Studies. 53: 89-94. effort to develop validated desertification monitoring techniques in Northern Africa. Proceedings of the international symposium: “Satellite-based observation: a tool for the study of the Mediterranean basin”. CNES, Toulouse, France. 34 Remote sensing, a tool to monitor and assess desertification
    • Young people of the Mbour and Gouvieux IRD youth clubs studying vegetation in the environmental work camp of Mbour. Several transects made up of five 10m*10m quadrats have been identified. This picture shows a wooded savannah plot, Senegal. Dominique Lefèvre © IRDFloret C. and Pontanier R., 1982. L’aridité en Tunisie pré- Moroccan examplesaharienne. Climat, sol, végétation et aménagement. Travaux etdocuments de l’ORSTOM n°150, Paris. 544 p. Baudat J., 2003. Une approche spatiale pour la surveillance de la faune. Étude de cas au Sud du Maroc : la vallée de l’oued Mird.Floret C., Le Floc’h E. and Rambal S., 1982. Measurement and Collection scientifique Roselt/OSS, Contribution Techniquemodelling of primary production and water use in a South n°10, Montpellier, France. 68 p.Tunisian steppe. Journal of Arid Environments. 5: 77-90. Blondel J., 1995. Biogéographie, approche écologique etFloret C., Le Floc’h E. and Pontanier R., 1992. Perturbations évolutive. Masson, Paris. 297 p.anthropiques et aridification en zone présaharienne. L’aridité,une contrainte au développement. Éditions ORSTOM Publishers, Brignon C. and Sauvage Ch., 1963. Planche 6B « Carte des étagesParis: 449-463. bioclimatiques du Maroc ». In: Comité de Géographie du Maroc (Ed.) Atlas du Maroc. Rabat.Jauffret S. 2001. Validation et comparaison de divers indicateursde changements à long terme dans les écosystèmes méditerra- Burel F. and Baudry J., 2001. Écologie du paysage, concepts,néens arides. Application au suivi de la désertification dans le méthodes et applications. 2nd edition. Éditions Lavoisiersud tunisien. Doctoral thesis, University of Marseille III (in coope- Publishers,Tec. & Doc., France. 349 p.ration with IRD), France. 372 p. Collar N.J., 1996. Family Otididae (bustards). In: Lynx EdicionsKennedy P.J., 1989. Monitoring the phenology of Tunisian Publishers, Handbook of the Birds of the World. Vol.3. Spain. 820 p.grazing lands. Int. J. Remote Sensing. 10(4-5): 835-845. Forman R.T.T. and Godron M.,1986. Landscape ecology.Long G., Le Floc’h E., Pontanier R., Debussche G. and Lacaze B., J. Wiley (Ed.), New York. 619 p.1978. Contribution à l’analyse écologique des zones arides deTunisie avec l’aide des données de la télédétection spatiale. Gaucher Ph., 1991. On the feeding ecology of the HoubaraExpérience ARZOTU, rapport final 1975-1978. CEPE/CNRS Chlamydotes undulata undulate. Alauda, Rev. Int. Ornitho.Montpellier, CNES, INRAT Tunis and ORSTOM, Paris. 222 p. 59(2): 120-121.Pontanier R., M’Hiri A., Akrimi N., Aronson J. and Le Floc’h E.(Eds), 1995. L’homme peut-il refaire ce qu’il a défait ? John LibbeyEurotext, Paris. 480 p.To go deeper... 35
    • WebsitesEuropean and international organisations • LANDSAT• Centre for the Study of the Biosphere from Space (CESBIO) http://landsat.gsfc.nasa.govwww.cesbio.ups-tlse.fr • MODIS• EUropean organisation for the exploitation of www.ga.gov.au/acres/prod_ser/modisdata.htmMETeorological SATellites (EUMETSAT) • MSGwww.eumetsat.de www.esa.int/msg/pag0.html• French Scientific Committee • PARASOLon Desertification (CFSD) http://smsc.cnes.fr/PARASOL/Frwww.csf-desertification.org • RADARSAT 2• Institut de recherche pour le développement (IRD) www.radarsat2.infowww.ird.fr • SMOS• Laboratory of Zonal Geography for the Development (LGZD) www.esa.int/esaLP/smos.htmlwww.univ-reims.fr • Spot 4• Long-Term Ecological Monitoring http://spot4.cnes.frObservatories Network (ROSELT) • TERRAwww.roselt-oss.teledetection.fr http://terra.nasa.gov• Regional Training Centre for Agrometeorology andOperational Hydrology and their applications (AGRHYMET) Satellite image providerswww.agrhymet.ne• Regional Research Network on Global Change in the • ESA (European Space Agency)Mediterranean basin and subtropical Africa (Médias) http://earth.esa.int/imageshttp://medias.obs-mip.fr • EURIMAGE• Sahara and Sahel Observatory (OSS) www.eurimage.comwww.unesco.org/oss • GLCF (Global Land Cover Facility)• United Nations Convention http://glcf.umiacs.umd.eduto Combat Desertification (UNCCD) • Landsatwww.unccd.int www.landsat.org • Radarsat InternationalSpace agencies www.rsi.ca • Space Imaging• Canadian Space Agency (CSA) www.spaceimaging.comwww.space.gc.ca/asc • Spot Image• European Space Agency (ESA) www.spotimage.frwww.esa.int • VITO (VEGETATION)• French Space Agency http://free.vgt.vito.be(Centre National dÉtudes Spatiales, CNES)www.cnes.fr Education• Indian Space Research Organisation (ISRO)www.isro.org • Canada Centre for Remote Sensing• Japan Aerospace Exploration Agency (JAXA) www.ccrs.nrcan.gc.ca/ccrs/learn/learn_f.htmlwww.jaxa.jp • Institut national agronomique Paris-Grignon (INA-PG)• National Aeronautics and Space Administration (NASA) http://lacan.grignon.inra.fr/ressources/www.nasa.gov teledetection/vademecum.htm• National Space Development Agency of Japan (NASDA) • Institut national agronomique Paris-Grignon /www.nasda.go.jp/index_e.html National Institute for Agricultural Research (INRA) www.inapg.inra.fr/ens_rech/ager/ressources/Satellite programmes and missions supports/courteld/cours/couverture.htm • Remote Sensing Tutorial (NASA)• ADEOS II http://rst.gsfc.nasa.govwww.eoc.jaxa.jp/adeos2• ENVISAT ISIS Programme / Incentive for the Scientifichttp://envisat.esa.int use of Images from Spot system• ERShttp://earth.esa.int/ers http://medias.obs-mip.fr/isis• IRS-P4 (OCEANSAT) http://medias.obs-mip.fr/www/Reseau/www.isro.org/irsp4.htm Lettre/13/en/isis.pdf36 Remote sensing, a tool to monitor and assess desertification
    • Co lour supplement Camp of Fulani herdsmen in a reg (desert pavement) invaded by the sand of the previous erg (Inchirian period, 22000 BP). North Oursi, Burkina Faso. Jean-Claude Leprun © IRD
    • Example (page 13)Use of the “roughness” parameterThe Earth seen from a radar altimeter:land topography and ocean bathymetry © ESA/ERS This map is drawn out from data derived from the ESA (European Space Agency) ERS-2 European satellite).Example (page 14)Use of the “soil moisture”parameter in Europe This soil moisture map is derived from measurements performed by the ERS satellite (January 2000) over Europe. Data are expressed comparatively, in percentage, with 0% representing dry lands and 100% very wet lands. Source: kindly provided by the Institute of Photogrammetry and Remote Sensing, Vienna University of Technology Soil moisture in Europe – ERS satellite (January 2000)II Remote sensing, a tool to monitor and assess desertification
    • Example (page 15)Global scale vegetation indicesA global synthesis of vegetation indices made by EOworks from © CNES 2004NDVI Spot-Vegetation data (03/1999). The colour range Image distributed by VITOexpresses increasing index values, from yellow to green. Example (page 17) Biomass monitoring in the West Sahelian region Estimating above-ground plant biomass resorts to meteorological data combined with those derived from scatterometers (through modelling). In the Sahelian area subject of this study, biomass distribution is significantly different in 1994 (wet year) and 1997 (dry year). This method applied to the whole Sahelian belt shown in this figure has been sub- mitted to ground validation in the Gourma region (Mali). Above-ground vegetation production in the West Sahelian region (-18°E/18°E and 13.5°N/20°N) assessed by an ERS Colour scale ranges from 0 (off-white) to 3,000 kg (dark brown) satellite scatterometer regarding of dry matter per hectare. these two contrasting years. © CESBIO 2003 From Jarlan et al., 2003.Colour supplement III
    • Examples (page 16) Use of vegetation indices: a map of African drylands Vegetation indices (NDVI) for the first ten days of April 2004 are derived from NOAA/AVHRR satellite data. The higher the index, the more developed is the vegetation cover. Source: Data from NASA GSFC (Goddard Space Flight Center), GIMMS (Global Inventory Modelling and Mapping Studies). Map issued by the FEWS-NET (Famine Early Warning Systems Network) project of the USGS EROS (United States Geological Survey – Earth Resources Observation System) Data Centre. A research funded by USAID (U.S. Agency for International Development). Monitoring of grass biomass in the pastoral area of Burkina Faso Annual maps of grass biomass in the pastoral area of Burkina Faso in 1999 and 2000 (note the difference between the two annual situations).In the Sahel, annual grasses that make up a large part of pastoral Similar techniques may be applied to other semi-arid pastoral areasresources are the predominant vegetation cover. The AGRHYMET rich in annual plants, such as Central Asian regions.Centre (CILSS: Permanent Inter-state Committee for Drought Control in Monitoring vegetation north of the Sahara is more difficult becausethe Sahel) created in 1974 is currently monitoring grazing lands at the steppes of small low woody bushes prevail. Moreover, during aregional scale and is circulating information to national decision- large part of the year, vegetation is only weakly green, or even notmakers. This allows to determine grazing lands at risk, and if need be, at all during very dry periods. It is then quite impossible to useto issue a warning in order to reduce grazing in relevant regions and vegetation indices to estimate variations in vegetation cover.prevent their desertification. The state of pastoral resources is asses- © AGRHYMET Regional Centresed by estimating biomass from cumulated vegetation indices. These Source: AP3A (Early Warning and Agricultural Production Forecast) Project.data regarding the Sahelian region are also listed in the monthly AGRHYMET Regional Centre, Niamey, Nigerreport of the FAO (Food and Agriculture Organization of the UnitedNations) GIEWS (Global Information and Early Warning System). IV Remote sensing, a tool to monitor and assess desertification
    • Focus (page 24)Degradation due to a decrease in the vegetationcover of the Menzel Habib sandy steppe, Tunisia View at ground level R. Escadafal © IRD Normal state Degraded state Satellite images R. Escadafal © IRD MSS image 04/76 TM image 04/89 TM image 03/93 TM image 03/99 Four georeferenced and intercalibrated Landsat images of the Menzel Habib area (Tunisia) (1976, 1989, 1993, 1999) Richard Escadafal © IRDThe differences evidenced correspond to modifications in surfacestates at ground level. Each picture covers a 30x24 km area. • Landsat TM image dated March 1993• Landsat MSS image dated April 1976 In 1993, i.e. four years later, the impacts of exclosure areas clearlyThe Menzel Habib plain in 1976: bordered in the South-East and South- show as dark trapezium-shaped spots among sandy areas, while annualWest by mountains (in grey), a sandy steppe prevails (in beige in the crops intensify (red dotted lines).centre). Annual crops (in red) are located at the periphery and in • Landsat TM image dated March 1999depressions (bright red spots). In 1999, mobile sand areas prove to have completely receded and• Landsat TM image dated April 1989 the situation is apparently well controlled. The landscape appears toIn 1989, drought leads to a decrease in crops and to the striking be fully parcelled out in plots of various uses, and new exclosure areasexpansion of mobile sand (light yellow): the region is “desertified”. have been set up in the few remaining dune areas.Colour supplement V
    • Focus (page 25)Synthesis of sand invasion evolutionin the Menzel Habib area (Tunisia)between 1989 and 1999 R. Escadafal © IRDThis figure illustrates one of the syntheses obtained by analysing trends Satellite images thus evidenced that the environmental state didobserved in images classified over five dates (between 1989 and improve between 1989 and 1999. Efforts undertaken in order to fix1999) and showing sand-invaded surfaces (unfixed moving sand, the moving sand that had invaded the region during the previousdunes, etc.). Sand-invaded areas (mobile dunes) have decreased in decade were successful. Space-based monitoring allowed tofavour of fixed sand areas and farming lands. quantify these impacts over large expanses. Such good results were of course also observed locally on site.VI Remote sensing, a tool to monitor and assess desertification
    • page 28 Landscape map, Oued Mird observatory, Morocco. © ROSELT/OSS.From Baudat, 2003. Lambert conformal conic projection / South Morocco II. Merchich datum / 1880 Clarke ellipsoid (IGN). Map derived from two 15-m (visible/NIR) Aster images. Pictures taken on August 18th, 2001. J. Baudat / Mastère SILAT (GIS applied to land use planning) IRD-ROSELT/OSS, October 2003.Colour supplement VII
    • page 28 The colour supplement is printed on Cyclus Print chlorine-free de-inked recycled paper. Location of the houbara bustard habitat (in dark orange) in Oued Mird. Modified topographic base: 1:100,000 (DFCTT, 1968). © ROSELT/OSS. From Baudat, 2003.VIII Remote sensing, a tool to monitor and assess desertification
    • AbstractRemote sensing is a technique that enables to observe the radiation scattered or emittedby the Earth surface. Satellite-based remote sensing allows regular, repetitive, accurate CSF SFD Les dossiers thématiquesobservations of nearly the whole planet, at various spatial and temporal scales, in severalwavelength fields.Such observations render the nature, state, temporal and spatial variations of theproperties of the objects at the Earth surface. By way of example, water-covered areas,roughness, soil moisture, changes in the nature of land, density and phenological In the same series:evolution of the vegetation cover, sand winds, are information included in theseobservations. Nevertheless, these observations usually combine together, making them English & French editionsmore or less difficult to extract from the raw data transmitted by satellites. The science ofremote sensing consists in interpreting and processing the series of spatial and temporal Is combating desertificationimages in order to extract such parameters, qualitatively or quantitatively.Desertification is a phenomenon of irreversible land degradation. It results from complex an environmental global public good?processes linked to the coupled and joint evolution of natural and human-induced Elements of an answer…factors. The beginning, development and results of such processes are materialised byland surface states and their evolution. La lutte contre la désertification : un bien public mondial environnemental ?Remotely sensed data consequently include information that the science of remote Des éléments de réponse...sensing allows to partly extract with more or less accuracy. Such information coupled (M. Requier-Desjardins & P. Caron)with others are involved in various stages of the desertification process. Remote sensingprovides useful data; some of them are essential information impossible to collectotherwise (especially in terms of homogeneity and spatial coverage and/or temporal Remote sensing,monitoring) for early warning, monitoring the development of desertification a tool to monitor and assess desertificationphenomena and acknowledging a final state. Among others, remote sensing may allow todetermine the impacts of policies to combat desertification. However, because of the La télédétection : un outil pour le suivimentioned limits regarding the extraction of useful parameters and the part played by the et l’évaluation de la désertificationlatter in the processes concerned, remote sensing turns out to be a tool among others - (G. Begni, R. Escadafal,certainly a powerful one, but not a scientific, decisional or operational “miracle” solution. D. Fontannaz & A.-T. Nguyen)After presenting in detail the technique and science of remote sensing and how it allowsto monitor various elements of desertification processes, this brochure deals with the French editionsmost important and significant cases and brings both aspects together. Several keyparameters and processes are studied: roughness, albedo, surface temperature, moisture,vegetation indices on the one hand; vegetation cover monitoring, modifications in the Restauration des milieux dégradésland surface composition in dry environments, wind transportation on the other hand. en zones arides et semi-aridesExamples are developed: evolutions of specific sites, projects under way. Lessons taught (É. Le Floch & J. Aronson)by previous experiments are critically analysed, options for the future are designed. Biodiversité et désertificationKey words: Remote sensing, desertification, vegetation cover, land surface states, (A. Sarr)wind transportation, early warning, prevention policies, monitoring/assessmentRésumé Pastoralisme et désertification en zone subsaharienne (Ph. Lhoste & B. Toutain)La télédétection est une technique permettant l’observation du rayonnement diffuséou émis par la surface de la Terre. La télédétection par satellite permet des observationsrégulières, répétitives, fidèles, de la quasi-totalité de la planète, à divers pas d’espace et La révolution pastorale en Méditerranéede temps, dans plusieurs domaines de longueur d’onde. et son impact sur la désertificationCes observations traduisent la nature, l’état, la variation spatiale et temporelledes propriétés des objets présents à la surface terrestre. À titre d’exemple, l’étendue (A. Bourbouze)des surfaces en eau, la rugosité, l’humidité des sols et leur changement de nature,la densité et l’évolution phénologique du couvert végétal, les vents de sable, sont desinformations présentes dans ces observations. Biens, ressources naturelles et pauvreté dans les sociétés pastorales :Néanmoins, ces observations se combinent entre elles le plus souvent et il est plus ou quelles approches ?moins difficile de les extraire des données brutes transmises par les satellites. La science (A. Bourgeot)de la télédétection consiste à interpréter et traiter les séries spatiales et temporellesd’images afin d’extraire ces paramètres qualitativement ou quantitativement. Ladésertification est un phénomène de dégradation irréversible des terres. Elle est lerésultat de processus complexes liés à l’évolution conjointe et couplée de facteurs naturels Érosion éolienne et désertificationet anthropiques. Le départ de tels processus, leur développement et leur résultat, se (M. Mainguet & F. Dumay)traduisent dans les états de surface du sol et leur évolution.Les données de télédétection sont donc porteuses d’informations que la science de la Désertification et gestiontélédétection permet partiellement d’extraire, avec plus ou moins de précision. Ces des ressources en eauinformations, couplées avec d’autres, interviennent dans diverses phases du processus dedésertification. La télédétection apporte des informations utiles, parfois indispensableset impossibles à acquérir autrement (notamment en termes d’homogénéité et decouverture spatiale et/ou de suivi temporel), pour l’alerte précoce, le suivi du Impact socio-économiquedéveloppement de phénomènes de désertification et le constat d’un état final. Elle peut de la désertificationnotamment permettre de dresser des constats sur les impacts de politiques de lutte.Néanmoins, de par les limitations signalées sur l’extraction des paramètres utiles et lerôle de ces paramètres dans ces processus en jeu, la télédétection s’avère être un outil L’information environnementaleparmi d’autres, certes puissant, mais non une solution ‘miracle’ scientifique, pour l’aide à la décisiondécisionnelle ou opérationnelle.Après avoir présenté de manière détaillée la technique et la science de la télédétection,et comment elle permet de suivre certains éléments des processus de désertification, Changement climatiquece dossier s’attache aux cas les plus importants et significatifs, en rapprochant les deux et désertificationaspects. Certains paramètres et processus clés sont étudiés : rugosité, albédo, températurede surface, humidité, indices de végétation d’une part ; suivi de la couverture végétale,modification de composition des surfaces en milieu aride et transport éolien d’autre Arbres, arbustes et produitspart. Des exemples sont développés : évolutions de sites particuliers, projets en cours de forestiers non ligneuxdéveloppement. Des leçons d’expériences passées sont analysées d’un œil critique et desvoies pour le futur sont dessinées.Mots clés : Télédétection, désertification, couvert végétal, état des surfaces,transport éolien, alerte précoce, politiques de prévention, suivi/évaluationCover picture (photomontage):Landscape: Irrigation in arid environment, Tunisia - (J. Pouget © IRD)Man: Life of the Fulani ethnic group, Burkina Faso - (J.J. Molez © IRD)
    • Ministère délégué à la Recherche Secretariat of the United Nations 1 rue Descartes Convention to Combat Desertification 75231 Paris CEDEX 05 P.O. Box 260129 France Haus Carstanjen Tel.: +33 (0)1 55 55 90 90 D-53153 Bonn www.recherche.gouv.fr Germany Tel.: +49 228 815-2800 Ministère des Affaires étrangères www.unccd.int 20 rue Monsieur 75007 Paris Agropolis France Avenue Agropolis Tel.: +33 (0)1 53 69 30 00 F-34394 Montpellier CEDEX 5 www.diplomatie.gouv.fr France Tel.: +33 (0)4 67 04 75 75 Ministère de l’Écologie et www.agropolis.fr du Développement durable 20 avenue de Ségur With special contribution from: 75302 Paris 07 SP Centre National dÉtudes Spatiales France 18 avenue Édouard Belin Tel.: +33 (0)1 42 19 20 21 F-31401 Toulouse CEDEX 9 www.ecologie.gouv.fr France Tel.: +33 (0)5 61 27 31 31 Agence Française de Développement www.cnes.fr 5, rue Roland Barthes 75598 Paris CEDEX 12 With the cooperation of: France Laboratoire de Géographie Zonale Tel.: +33 (0)1 53 44 31 31 pour le Développement www.afd.fr (Université de Reims Champagne-Ardenne) 57 rue Pierre Taittinger F-51096 Reims CEDEX France Tel.: +33 (0)3 26 91 81 62 www.univ-reims.frHOW TO CONTACT US CSFD Comité Scientifique Français de la Désertification Agropolis International Avenue Agropolis F-34394 Montpellier CEDEX 5 France Tel.: +33 (0)4 67 04 75 44 Fax: +33 (0)4 67 04 75 99 csfd@agropolis.fr www.csf-desertification.org